Max Children Issue and SA timeouts

Jason Burzenski jburzenski at AMERICANHM.COM
Fri Apr 23 19:24:39 IST 2004


> You should upgrade SA. It not unlikely it's part of the
> problem and 2.63 will catch more spam anyway.

I haven't had time to read up on converting my bayes from 2.55 to 2.63 and I
really need to test the conversion before I roll 2.63 out into production.
I tend to rely heavily on my bayes for effective filtering.  But, since I
keep getting pointed in this direction, I will escalate my priority on
getting this upgraded.  (I was hoping to wait until SA3/MS4.3 was stable,
but that will probably take awhile).

> Did you get your Bayes problems fixed? I vaguely remember
> that Bayes was mentioned when you posted the other week. The
> expire problem can cause SA to time out real bad. How does
> your Bayes directory look? You should only have three files
> there, one journal, one seen and one tokens.

When I saw the SA timeout count at 0 on the box with the bayes problem, I
was in even less of a hurry to fix it.  Here is my bayes directories on both
boxes:

Box1 (sporadic MS process count, *lock.box* files disappear and reappear
constantly)
drwxr-xr-x    2 root     root         4096 Apr 18 00:25 1
-rw-------    1 root     root       339929 Apr 23 14:22 bayes_journal
-rw-------    1 root     root           27 Apr 23 14:18 bayes.lock
-rw-------    1 root     root          216 Apr 23 14:22
bayes.lock.box1.13716
-rw-------    1 root     root          189 Apr 23 14:22
bayes.lock.box2.13734
-rw-------    1 root     root          108 Apr 23 14:22
bayes.lock.box3.13738
-rw-------    1 root     root         4720 Apr 23 14:22 bayes_msgcount
-rw-------    1 root     root     14139427 Apr 23 11:25 bayes_seen
-rw-------    1 root     root     11071488 Apr 23 14:18 bayes_toks
-rw-------    1 root     root      6844416 Apr 23 14:18 bayes_toks.new
-rw-r--r--    1 root     root         1218 Dec  8 10:40 user_prefs

Box2 (stable MS process count, never see *lock.box* files)
-rw-------    1 root     root       101409 Apr 23 14:24 bayes_journal
-rw-------    1 root     root           26 Apr 23 14:24 bayes.lock
-rw-------    1 root     root         1928 Apr 23 14:24 bayes_msgcount
-rw-------    1 root     root     14225408 Apr 23 14:23 bayes_seen
-rw-------    1 root     root     123088909 Apr 23 14:24 bayes_toks
-rw-------    1 root     root     18644992 Apr 23 14:20 bayes_toks.new
-rw-r--r--    1 root     root         1218 Dec  8 10:58 user_prefs


Looks like my toks are getting pretty big on box2.  Could this be party of
the problem?

Thanks again for all of your help!

-------------------------- MailScanner list ----------------------
To leave, send    leave mailscanner    to jiscmail at jiscmail.ac.uk
For further info about MailScanner, please see the Most Asked
Questions at    http://www.mailscanner.biz/maq/     and the archives
at    http://www.jiscmail.ac.uk/lists/mailscanner.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.mailscanner.info/pipermail/mailscanner/attachments/20040423/e157920b/attachment.html


More information about the MailScanner mailing list